A Model-Based Trajectory Planning Method for Robotic Polishing of Complex Surfaces

Off-line programming of the polishing tool trajectory for complex workpieces is challenging due to the nontrivial material removal model and the polishing accuracy requirement. Current tool trajectory planning methods are mainly developed for some simple surfaces but cannot handle the increasingly complicated industrial parts, such as the wheel hubs. This article first develops a numerical contact mechanics model for the point-sampled complex workpieces. The contact pressure distribution and the material removal depths on the workpiece point cloud can be predicted efficiently. A novel high-priority subregion searching algorithm is developed to track the most-worth-polishing workpiece points. By selecting the path pattern as direction-parallel, the path direction, tool dwell times, and the path spacings inside each extracted subregion are optimized to minimize the deviation from the desired material removal depths. The effectiveness of the proposed method is verified by performing disk polishing simulations on workpieces with different shapes. A robotic polishing experiment is also conducted on a wheel hub. Both simulation and experimental results show that reasonable tool trajectories can be generated on the workpiece, and the desired material removal depths can be achieved. Note to Practitioners—In robotic polishing industries, it is crucial to plan the tool trajectory (tool path and feed velocity) to achieve desired material removal depths on the workpiece surface, which means high surface quality. In this article, a model-based tool trajectory planning method for robotic polishing of complex surfaces that are represented by the point cloud form is presented. The advantage of using the point cloud is that workpiece surfaces with varying curvatures and complex features, e.g., grooves and holes, need not be expressed explicitly. The proposed method generates high-priority subregions according to the updated material removal distribution dynamically. In this work, the polishing path pattern is chosen as direction-parallel. Based on an efficient numerical contact mechanics and material removal model, the path locations and the tool dwell times inside each subregion are optimized to minimize the deviation between the actual and the desired material removal depths. When the desired material removal depths are attained in an extracted subregion, the algorithm finds the next high-priority subregion until the whole workpiece is well polished. The trajectory planning method can be integrated into an industrial robot with the force-control module. Future work is to integrate the roughness model into the tool trajectory planning method.